This proposal has five goals. 1) To assess the risks, if any, of caffeine use in pregnancy on fetal growth as assessed by low birth weight, preterm delivery and intra-uterine growth retardation. Almost without exception, the existing literature is based on secondary analyses of data collected for other purposes. This may account for lack of agreement among these studies. 2) The role of decaffeinated coffee on the same reproductive outcomes will be evaluated. This beverage is being increasingly consumed by pregnant women but there are no studies of its effects on reproductive outcomes. Because of difficulties in disentangling the highly correlated exposure to caffeine and coffee, any reproductive effects attributed to caffeine may be associated with other agents in coffee. 3) We plan to measure interactive effects of caffeine and cigarette smoking in detail using biologic markers. Caffeine and cigarette use are highly correlated but also interact in important ways. Caffeine appears to be metabolized more rapidly by cigarette smokers and the independent effect of either exposure on fetal growth can only be assessed with careful measurement of both. The study will assess maternal urinary cotinine and caffeine at selected points in pregnancy, as well as in amniotic fluid and newborn urine. 4) Caffeine metabolism may be strongly influenced by pregnancy itself, metabolism slowing in the later stages of pregnancy. The study will evaluate urinary caffeine and questionnaire consumption data at selected intervals throughout pregnancy to study these effects in detail. 5) Questionnaires collecting caffeine consumption data, validated using biologic markers of exposure and metabolism, will be available for future studies. The study will use a nested cohort design in which subgroups of a large cohort (n=3,000), selected randomly, are monitored in detail for biologic markers of interest. Caffeine, decaffeinated coffee and cigarette exposure are all common exposures in pregnancy, thus even relatively small increased risks will have large public health effects. The study will detect risks ranging from 1.75 to 2.0 (a=0.05, B=0.10). Preterm delivery and intra-uterine growth retardation are major health problems in their own right but also among the strongest predictors of infant mortality and long-term disability.